Of the above two methods, that in which an ammonium oxalate solution is used is the more delicate, although it cannot be employed quantitatively, owing to the oxidation of the metal that takes place. An addition of 1 grm. of ammonium oxalate to the suspected solution is sufficient. _Copper_.--0.00005 grm. of copper can be very readily detected by electrolyzing an acid solution in the usual way. A spiral of platinum wire is employed as the cathode, and the presence of the metal confirmed for by dissolving it in a little nitric acid, diluting with water and adding potassium ferrocyanide. To detect these metals in cases of poisoning, the organic matter with which they are associated must first be destroyed in the usual way by means of hydrochloric acid and potassium chlorate, and the precipitates obtained in the ordinary course of analysis, then subjected, at suitable stages, to electrolysis. As the solutions thus obtained will be still contaminated by some organic matter, it is necessary to pass the current for a longer time than indicated above. On the other hand, _urine_ can be tested directly for these poisons. The presence of mercury or of copper may be detected by acidifying the urine with 2-3 c.c. of nitric acid (conc.), and electrolyzing as described. 0.0001 grm. of metal in 30 c.c. of urine can be detected thus, or 1 part in 300,000 of urine. Lead does not separate well as peroxide from urine, but if ammonium oxalate be added, and the lead deposited as metal, the reaction is quite as delicate as in aqueous solution, and 0.0001 grm. of lead can be thus detected. With antimony it is advisable to precipitate it first as sulphide, but it can be detected directly, though not so satisfactorily, by acidifying the urine with 2-3 c.c. of sulphuric acid (dil.), and electrolyzing with a current of 1-5 to 2 c.c. In this case also it is precipitated as metal upon the cathode (cp. Chittenden, Proceedings Connecticut Acad. Science, Vol. 8). In the presence of urine it is advisable to continue the passage of the current for about twice the time required in the case of aqueous solutions. That an approximately quantitative result can be obtained under the above conditions was shown in several cases in which deposition of 0.001 grm. of metal was confirmed with considerable accuracy, the spiral or foil being weighed before and after the experiment.